Quantification of Volatiles from Technical Lignins by Multiple

Subscriber access provided by Stockholm University Library

Article

Quantification of Volatiles from Technical Lignins by MHS-SPME-GC-MS Matthias Guggenberger, Antje Potthast, Thomas Rosenau, and Stefan Böhmdorfer ACS Sustainable Chem. Eng., Just Accepted Manuscript • DOI: 10.1021/ acssuschemeng.9b00630 • Publication Date (Web): 26 Apr 2019 Downloaded from http://pubs.acs.org on April 27, 2019

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 17 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Sustainable Chemistry & Engineering

Quantification of Volatiles from Technical Lignins by MHS‐SPME‐GC‐MS Matthias Guggenberger,† Antje Potthast,† Thomas Rosenau,†,‡ Stefan Böhmdorfer†,* †

University of Natural Resources and Life Sciences, Vienna (BOKU), Department of Chemistry, Institute for Chemistry of Renewable Resources, Konrad‐Lorenz Straße 24, 3430 Tulln, Austria

‡

Johan Gadolin Process Chemistry Centre, Åbo Akademi University, Porthansgatan 3, Åbo/Turku FI‐20500, Finland *[email protected]

Abstract Industrial lignins comprise a mixture of substances, including volatile, low‐molecular weight compounds. In material applications of lignins, these volatiles contribute to the malodor of the finished product. We developed a method based on SPME‐GC‐MS to assay qualitatively and quantitatively the volatiles emitted from lignin samples. Substances were identified by mass spectra and retention indices, while quantitation was achieved by multiple headspace sampling (MHS). Guaiacol and dimethyl disulfide were calibrated as representative compounds for the most prominent substance classes. The method was validated and gave good recovery – ranging from 89 to 123% for dimethyl disulfide and 90 to 105% for guaiacol – a measurement range of several dozen nanogram to few microgram, which can be extended by adjusting the sample amount, and limits of detection of 86 ng for dimethyl disulfide and 25 ng for guaiacol. Sample preparation is limited to weighing of the sample into a headspace vial and requires no consumables or auxiliaries. The entire analytical workflow was automatized, including the necessary data evaluation, which combines the outcome of repeated analyses of the same sample. The concentrations of guaiacol in four representative lignin samples ranged from 0.4 ppm to 1200 ppm, while dimethyl disulfide was detected only in a single sample.

Keywords Multiple Headspace Solid‐Phase Microextraction, Volatiles, Quantification, Lignin, Lignosulfonate, Renewable Resources, Guaiacol, Dimethyl disulfide

Introduction Lignin is one of the most abundant biogenic polymers. It is found in vascular plants and is derived from monolignols – a set of differently substituted hydroxycinnamyl alcohols – via random radical polymerization.1 Technical lignins are produced on a megaton scale by the pulp and paper industry. They offer a two‐fold benefit as raw material: 1) there is a major economic benefit in using technical lignins for higher value, material applications; 2) as renewable, plant‐based raw materials they inherently reduce the emission of fossil carbon dioxide.2,3 One caveat in the development of lignin applications is its odor, which is caused by small molecules originating from the actual lignin polymer – e.g. guaiacol – and from pulping chemicals (e.g. dimethyl disulfide ‐ DMDS).4 Page 1 of 17

ACS Paragon Plus Environment

ACS Sustainable Chemistry & Engineering 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 2 of 17

Nevertheless, literature on this topic is scarce, so that the picture on volatile substances released by technical lignins is still incomplete. One of the earliest publications solely dedicated to lignin volatiles was elaborated in 2008.5 Therein, Rocha et al. report the identification of volatiles from dioxane and organosolv lignins as well as their discrimination by means of their corresponding volatile profiles only. The identification was based on SPME‐GC‐MS (solid‐phase microextraction‐gas chromatography‐mass spectrometry) and the discrimination was based mainly on the use of a proprietory aroma sensing system combined with principle component analysis (PCA). Although a first description of lignin volatiles was achieved, no absolute quantitative results were presented. Another publication, which mainly dealt with the effect of enzymatic treatments on the odor profile of kraft lignins, also reported quantitative results for guaiacol, a well‐known lignin volatile.4 These were measured via thermodesorption (TD) at elevated temperatures (150°C and 190°C) coupled to GC‐MS, whereby quantification was achieved by external calibration. As the focus of this work was centered on lignin treatments, not many details of the quantitative analysis were shown. Moreover, an influence of matrix effects on the described approach might be present. A study on patterns of volatiles at even higher temperature was conducted by Voeller et al. Thermal carbon analysis in combination with thermodesorption‐pyrolysis‐GC‐MS (TD‐Py‐GC‐MS) was used to examine compounds evolving at temperatures from 200°C up to more than 800°C. At these temperatures, lignin degradation adds phenolic dimers and oligomers to the group of emitted substances.6 Kouisni quantified kraft lignin sulfur volatiles via external calibration. This approach was based on headspace sampling and GC with a sulfur specific detector. Also here, not many details on the quantification protocol were mentioned.7 Volatiles released from soda lignin and a blend thereof with polyamide were investigated with TD‐GC‐MS.8 One analyte, namely formaldehyde, was quantified by derivatisation followed by HPLC analysis. Here only the amounts of formaldehyde were given in absolute numbers; all GC‐MS derived values were expressed as toluene equivalents. As the examples above show, the analysis of lignin volatiles has been tackled from different directions. However, none thereof present a generally applicable method for absolute quantification of these volatile substances. With our work, we try to fill this gap and provide a reliable and facile method for the absolute quantification of a variety of lignin volatiles. The heart of this approach is SPME‐GC‐MS together with a special sampling scheme called multiple headspace extraction (MHE). This combination is also known as multiple headspace sampling ‐ solid‐phase microextraction (MHS‐SPME).9 The basic concept underlying MHE and therefore also MHS‐SPME is that by repeatedly analyzing one sample under fixed conditions, the analyte is removed repeatedly and depleted resulting in an exponential decrease of the analyte signal area. This decrease can be described by the following equation: 𝐴

𝐴 ∙𝑒

∙

1

with 𝐴 , the area of the analyte signal after the 𝑖 th extraction, 𝐴 , the area of the analyte signal after the first extraction and the constant 𝑞′, which does not have a tangible physical meaning.10 According to that, infinitely many extraction steps would exhaustively extract all the analyte from the sample and summation of all areas, 𝐴 , will lead to a total area, 𝐴 , relating to the total amount of analyte in the sample. Fortunately, the sum of 𝐴 is in the form of a geometric progression that converges to a limit. Hence 𝐴 can be inferred by calculation: 𝐴

𝐴

𝐴 1

𝑒

Page 2 of 17


2

Page 3 of 17 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


The only unknown parameter of this equation, 𝑞′, is obtained by linear regression. To this end, the natural logarithms of measured 𝐴 values are plotted against 𝑖 1 and fitted according to: ln 𝐴

𝑞 ∙ 𝑖

1

ln 𝐴

3

whereby the slope of the regression line corresponds to 𝑞′. From this value together with 𝐴 the total area can be calculated and correlated to the actual sample amount via a conventional calibration curve. 9,10 Although this procedure may seem complicated and tedious, it has a big advantage: due to the fact that the total amount of analyte is measured via its stepwise depletion, matrix effects that might interfere with equilibration in the headspace are compensated.9,10 Furthermore, it was shown that MHS‐SPME is even applicable in situations where equilibrium has not been reached, as long as all the measurement parameters are kept constant throughout the extractions.9,11 Since matrix effects are overcome by MHS‐ SPME, it is predestined for samples with volatiles in complex matrices that are unsuited for the standard addition technique. This is true for many solid samples. Examples for such samples that have already been analyzed by MHS‐SPME are soil,12 polyamide,13 cork stoppers,14 rosemary extract,15 sausages,16 tomatoes,17 mushrooms,18,19 spices20 and ink on paper21. The following paragraphs give a detailed picture on the method itself as well as on its development. Guaiacol and DMDS, two known lignin volatiles4,7 were selected for the analysis of four representative technical lignins in order to illustrate the workings of the analysis method.

Materials and Methods

Sources and Purities of the Used Chemicals

Ethyl acetate, analytical grade, from Merck (Darmstadt, Germany); dimethyl disulfide (DMDS), ≥99%, from Sigma Aldrich (St. Louis, Missouri, U.S.A.); glycerol, analytical grade, from Sigma Aldrich (St. Louis, Missouri, U.S.A.); guaiacol, 98%, from ABCR (Karlsruhe, Germany); iso‐propanol, HPLC grade, from Roth (Karlsruhe, Germany). All chemicals were used without any further purification.

Standard Solutions and Samples

Calibration and validation standards were prepared by dissolving 50 mg of the pure substance (guaiacol or DMDS) in 5 g of glycerol. 50 mg of this stock solution were diluted with 10 g of glycerol to obtain the working solution. As MHS‐SPME measures not the concentration but the total amount of analyte in the sample, the calibration was done by weighing different amounts of working solution (typically between 1 and 100 mg) into 10 mL headspace vials with magnetic screw caps and septum (silicone/PTFE; thickness: 1.0 mm) and subsequently analyzing them. Lignin samples were prepared similarly by weighing the desired amount of lignin powder into the headspace vial. In this study, four different technical lignins were examined: Commercial Indulin kraft lignin (sample I, stored at ambient conditions for three years in its original paper bag) another technical kraft lignin (sample Page 3 of 17



II, stored for a few month in a closed plastic container in the refrigerator), a commercial ammonium lignosulfonate (sample III, in ambient storage for a few years), and a beech organosolv lignin (sample IV, in ambient storage for a few years).

GC‐MS Conditions

GC‐MS measurements were carried out with an Agilent 7890B GC system coupled to a 5977A quadrupole MS Detector. The GC was equipped with two inlets, a multimode inlet connected to a short (around 20 cm) and plugged piece of fused silica tubing protruding into the GC oven, and a split/splitless inlet connected to an Agilent J&W VF Wax GC column (30 m long; inner diameter: 0.25 mm; film thickness: 0.25 µm, Santa Clara, California, U.S.A). Straight SPME liners from Agilent (0.75 mm inner diameter) were used in both inlets. The multimode inlet was held constant at the conditioning temperature (300°C for carboxene polydimethylsilane (PDMS) fibers). It was used to condition SPME fibers directly before a measurement series. The split/splitless inlet was operated at 280°C in splitless mode with 3 mL min‐1 septum purge flow. Desorption of the fiber was performed for 1.5 min in this inlet. After that, the fiber was kept in the split/splitless inlet for further 13.5 min with the split vent opened in order to clean the fiber for the next measurement. The GC was operated with helium at 0.9 mL min‐1. During a measurement, the GC oven was kept at 40°C for 1 min, followed by a ramp to 240°C at 10°C min‐1. The final temperature (240°C) was held for 4 min. The column effluent was transferred to the MS detector at 280°C. The ion source (electron impact, 70 eV) was kept at 230°C, while the mass analyser was set to 150°C. It was operated simultaneously in SIM (79 m/z; 94 m/z; 95 m/z; 124 m/z; 125 m/z) and Scan mode in order to get the full chromatogram in addition to characteristic selected ions corresponding to the desired volatiles. The mass range in scan mode was 29‐450 m/z. The entire GC‐MS system was controlled with MassHunter (B.07.02.1938) from Agilent (Santa Clara, California, U.S.A).

MHS‐SPME Conditions

Sample handling and SPME manipulation was performed with a CombiPAL autosampler (CTC Analytics AG) mounted onto the GC‐MS system. It was equipped with two sample trays, and a heatable agitator. The software tool Chronos (4.3.1.3837, Axel Semrau®, Sprockhövel, Germany) was used to control the autosampler and initialize GC‐MS runs. SPME fibers (fiber length: 10 mm, needle outer diameter: 23 and 24 ga equivalent to 0.64 mm and 0.57 mm respectively) and holders were obtained from Supelco (St. Louis, Missouri, U.S.A.). Right before use, the fibers were conditioned according to the supplier’s recommendations. MHS‐SPME measurements were carried out as follows: 10 mL headspace vials containing the sample were equilibrated at 40°C in the agitator (500 rpm) for 5 min. Then the fiber was placed in the vial for 30 min to extract the volatiles. This corresponded to the GC run time and ensured the removal of a sufficiently high amount of analytes from the sample to effect a decrease in peak area from one extraction step to the next, allowing a meaningful fit in the determination of 𝐴 . The extraction step was done in the heated agitator, however without shaking. Directly after extraction the fiber was desorbed in the split/splitless inlet and the GC‐MS analysis conducted as described in the previous section. Up to seven consecutive extraction steps were conducted for each sample. For calibration, validation, and final sample analysis five extractions Page 4 of 17


Page 4 of 17

Page 5 of 17 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


were performed. Between two samples a blank vial, containing either just air or a quantity (typically 100 mg) of neat solvent in the case of calibration solutions, was extracted and measured once.

Data Processing

Chromatograms were examined with Chemstation (F.01.01.2317, Santa Clara, California, USA) and OpenChrom (1.2.0 – Alder, Lablicate GmbH, Hamburg, Germany). Volatile compounds were identified via comparison with the NIST11 MS spectra library. For confirmation, retention indices calculated following the recommendations of the IUPAC gold book were compared to those provided on the Pubchem website.22,23 For quantification, the data was automatically processed with Origin (Origin 2018, OriginLab Corporation, Northampton, Massachusetts, U.S.A.) with a custom Origin C script. First, the spectra were exported from Chemstation in the AIA/ANDI format. Then, all chromatograms of one sample were identified and sorted into folders. For each chromatogram of one sample, a single ion trace of the desired mass to charge ratio (guaiacol: 81 m/z ±0.2 m/z; DMDS: 79 m/z ±0.2 m/z or 94 m/z ±0.2 m/z) was extracted from the full scan chromatograms. Next, peak areas were determined by integrating the single ion traces within a predefined retention time interval (guaiacol: 13.7‐14.0 min; DMDS: 3.8‐4.9 min). The natural logarithms of these peak areas were then fitted linearily and subsequently the total area, 𝐴 , was calculated according to the formula stated in the introduction. Finally, the total areas were used to calculate with a spreadsheet the amount of analyte present in the sample using an external calibration.

Validation

Validation was performed according to ICH guidelines.24 The actual validation process involved measuring glycerol solutions with known amount of neat analytes (see Standard Solutions and Samples). GC‐MS analysis and data processing was carried out as described in previous sections. For guaiacol, three sample series of five different analyte levels in triplicate were measured. For DMDS two sample series of five different analyte levels in triplicate were analysed. Measurements were conducted in random order. Random numbers generated by www.random.org were used for randomization.25 The resulting total areas were used to test the linear relationship of the total area and weighed analyte amount by the F‐test at a confidence level of 95%. Further, recoveries were determined individually for each calibration level by calculating the measured amounts of analyte at one analyte level by using the remaining results of the series for calibration. Limit of detection (LOD) and limit of quantification (LOQ) were calculated according to the signal‐to noise approach. To render this method applicable to MHS‐SPME, it was assumed that for an analyte present in a very small amount, one extraction would be sufficient to extract it completelyfrom the sample. The integral of the single analyte peak would then represent the total area which is used to calculate the analyte’s amount via a calibration curve. Hence, together with the ICH definition that LOD refers to a signal‐to‐noise ratio (SNR) of 3:1 and LOQ to a SNR of 10:124, one can define the areas to calculate LOD and LOQ as follows: 𝐴

𝐿𝑂𝐷

𝐴

⋅3

4

𝐴

𝐿𝑂𝑄

𝐴

⋅ 10

5

Page 5 of 17



Page 6 of 17

𝐴 𝐿𝑂𝐷 and 𝐴 𝐿𝑂𝑄 are the total areas of the LOD and the LOQ, respectively, and 𝐴 is the area obtained from blank samples (n=31; one blank sample contained 100 mg of neat glycerol) using the selected ions (guaiacol: 81 m/z; DMDS: 94 m/z) and integration parameters (guaiacol: 13.7‐14.0 min; DMDS: 3.8‐4.9 min) of a specific analyte. These total areas are then converted to LOD and LOQ amounts using the established calibration.24 The statistical analysis was aided with the NeoLiCy® software suite (NeoLiCy® version 2.1.3.0; QuAnalChem).

Results and Discussion The method was developed with the aim to analyze and quantify volatiles emerging from lignin. GC is the most appropriate chromatographic method for volatile compounds. The first step in method development was to identify a sampling system that would transfer the volatiles, which are evidently present in the gas phase into the GC for analysis. First experiments made use of headspace (HS) sampling, an approach that transfers the gas volume over the sample directly into the GC. However only under rather harsh equilibration conditions (80°C) some volatiles could be detected, albeit with a very low sensitivity. Therefore SPME, a technique which adsorbs and thus concentrates the analytes onto a fiber, was tested. This improved sensitivity and increased peak intensity by several orders of magnitude; with SPME sampling several signals were detected that were not observable with HS sampling (see Figure 1). Additionally, equilibration at elevated temperatures was not necessary for SPME‐GC‐MS. Consequently, SPME was chosen as the basis for all further development steps.

Figure 1: Comparison of chromatograms resulting from HS‐GC‐MS (plot A) and SPME‐GC‐MS (plot B) of samples containing sample I (mA= 0.503 g; mB= 2.014 g). HS‐GC‐MS was carried out on an Agilent 6890N GC with a 5975B MS detector and a 7697A headspace sampler. For separation, an HP‐5ms column was used (length: 30 m; inner diameter: 0.25 mm; film thickness: 25 µm). Headspace samples (1 mL) were introduced to the inlet in splitless mode after 5 min of equilibration at 80°C. SPME‐GC‐MS was performed as described in the Material and Methods, with a 30 min extraction period at room temperature. The chromatograms were scaled according to the amount of sample. Note the different scales of the Y‐axes. Page 6 of 17


Page 7 of 17 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Fiber Selection

Next, the SPME fiber coating which would extract the most compounds was selected by testing different fiber materials on the same lignin sample. For this sample, a high amount of lignin was used (2.014 g) in order to keep the amount of volatiles close to saturation over several measurements. As can be seen from Figure 2, all polydimethylsiloxane (PDMS) containing fibers perform satisfactory. Compared to the other fiber types they yield more peaks at higher intensities. For our purposes, carboxene‐PDMS fibers were chosen, as they gave the highest abundance of peaks, predominantly in the region of low molecular weight volatiles (i.e.: early in the chromatogram). Two substances were always present regardless of the fiber used: acetic acid at 9.7 min and guaiacol at 14.7 min. Guaiacol is a typical decomposition product of lignin’s polyphenolic structure4, and can therefore be expected to prevail in most lignin samples. It is more suitable for GC analysis than acetic acid, which is also reflected in the two compounds’ peak shapes (see Figure S1). Guaiacol was therefore selected for quantitation. The quantitative analysis of one compound establishes a reference that can be used to quantify all other components by GC‐FID (flame ionization detection) without additional calibrations if two prerequisites are met: the component must contain carbon to be detectable by FID, and it must be identified to calculate a response factor relative to guaiacol, in our case by MS and retention indices.26,27 Additionally, a quantification for DMDS was established to investigate the suitability of the method for molecules that have a high sulfur content, which can readily arise from kraft lignins.4,7

Page 7 of 17



Figure 2: Comparison of the performance of different SPME fiber materials. The sample was Indulin kraft lignin (2.014 g). Extraction at room temperature was performed for 30 min. Peaks marked with an asterisk do not originate from the sample but from the fiber.

MHE‐SPME: Initial Trials

Using sample II – the technical kraft lignin – analyses were performed to verify the suitability of MHE for lignin samples. As predicted by theory, chromatographic peaks decreased with successive extractions confirming the adequacy of the chosen extraction time, and the decline of peak areas showed the desired exponential behavior: the ln 𝐴 versus 𝑖 1 plot behaved as expected and could be fitted very well 0.9930 for the guaiacol peak of Figure 3). Linearity of this graph is a prerequisite to calculate linearly (𝑅 the total area of a component reliably.28 In addition to that, it was found that extraction at controlled temperatures as well as a short equilibration before SPME‐extraction was necessary to prevent influences by fluctuations in ambient temperature.

Page 8 of 17


Page 8 of 17

Page 9 of 17 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Figure 3: MHE analysis of a technical kraft lignin (sample II; m= 5.6 mg). A: Selected ion chromatograms of the guaiacol peak (81 m/z) of seven successive extractions. B: The corresponding integrated peak areas, 𝐴 , which decay in an exponential fashion during the course of the extractions, 𝑖. C: The natural logarithms of the peak areas, ln 𝐴 , versus extraction number 𝑖 1 as well as the resulting linear regression (ln 𝐴 0.9930). This slope is used together with the 0.633 ⋅ 𝑖 1 23.211; dashed line; 𝑅 peak area of the first extraction, 𝐴 , to calculate the Total Area, 𝐴 , using Equation 2.

Sample Amount

The amount of sample used for one set of extractions is an important parameter in MHE. While using too little obviously results in too low signals for reliable analysis (see Figure 4B), too much sample is also detrimental. The detected MS signal must be in the linear, non‐saturated range to determine meaningful peak areas. To prevent saturation, a single ion was chosen as the basis for integration and not the total ion current. While this ion (for example 81 m/z for guaiacol) was selected to have a high intensity, it was not necessarily the base peak in the molecule’s spectrum (109 m/z for guaiacol), as illustrated in Figure S2. Furthermore, we observed that for some samples too much material in the vial resulted in fluctuating peak areas during the extraction steps. The reason for that might be saturation of the fiber by other volatiles present in the sample’s headspace.28 This caused a non‐exponential decay that could not be fitted linearly, which in turn prohibited the calculation of a total area for those samples (Figure S3). To avoid these phenomena, sample masses of 15 mg for sample II and 100 mg for glycerol standards were not exceeded. Samples I and III allowed up to 1 g of sample mass. Masses up to 200 mg were applicable to sample IV.

Impact of the Number of Extractions

The necessary number of measurements for a single sample has a substantial impact on the analysis time per sample. Thus, we performed a set of experiments to determine the number of extractions cycles that is needed for reliable quantitation, and to clarify the influence of the extraction number on the final results. Comparable amounts of guaiacol and DMDS standards (around 2 µg) were measured with seven extractions, and the obtained datasets were processed repeatedly, adding different numbers of extractions to the linear fit. Two different scenarios were observed (denoted A for guaiacol and B for DMDS, see Figure 4). As can be seen from the obtained chromatographic peaks (top row), A is not extracted exhaustively, while compound B is depleted already with the third of five extractions. The linear fits of compound A are consistent, while they deviate substantially for compound B with increasing number of extractions included in the fit (middle row). The mean deviation from the actual analyte amount was Page 9 of 17



Page 10 of 17

calculated from three repeated measurements per analyte. As a consequence, the amounts calculated from these fits have a lower measurement uncertainty for compound A, remaining practically unchanged after five extractions, while the calculated amount for compound B is becoming increasingly biased (bottom row). Therefore, the quantitation must exclude all analyses exceeding analyte exhaustion, as this would increasingly distort the fit. In this case, a very low number of extractions will give results with a lower uncertainty than with a higher number of extractions. It might also be advisable to increase the sample amount to resolve this issue. For analytes that followed the expected logarithmic decrease of peak areas, five extractions were considered to be ideal.

Figure 4: Relationship between the number of extractions, the extraction behaviour of the analyte and the resulting measured amounts of two samples: A (guaiacol, which follows the expected logarithmic decrease of peak areas) and B (DMDS, which is exhaustively extracted after the second or third extraction). Top row: Chromatographic peaks of the two analytes. Middle row: ln 𝐴 plots with regression lines obtained from fitting with the inclusion of varying numbers of extractions. Bottom row: Calculated amounts (average from three measurements). The gray boxes visually indicate the range of the obtained values. Page 10 of 17


Page 11 of 17 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Validation

Calibration curves for guaiacol and DMDS are depicted in Figure 5, the results of the validation in Figure 6. For both analytes, a linear dependence of recovered mass versus actual analyte mass was found with a slope of approximately 1 and an intercept close to 0. This indicates that the measured values can be expected to reflect the true values closely. Depending on the analyte amount, recovery ranged from 89% to 123% for DMDS and 90% to 105% for guaiacol, when measuring amounts above the limit of quantitation.

Figure 5: Calibration curves of the two investigated volatiles. Panel A shows the curve for the 81 m/z ion of guaiacol the one for the 94 m/z ion of DMDS. The following LOD and LOQ values were found for guaiacol: 𝐿𝑂𝐷

25 𝑛𝑔

5.3 𝑛𝑔; 𝐿𝑂𝑄

86 𝑛𝑔

0.2 𝑛𝑔; 𝐿𝑂𝑄

35 𝑛𝑔

18 𝑛𝑔

6

And for DMDS: 𝐿𝑂𝐷

88 𝑛𝑔

0.6 𝑛𝑔

7

As can be seen from Figure 6 (bottom), the relative standard deviation (RSD) is rather stable over the calibration range (11% for guaiacol, 13‐30% for DMDS) and increases sharply in the proximity of the LOQ. For guaiacol, the LOQ based on noise is at an analyte mass that is already affected by a high measurement error and should therefore be set at a higher mass, namely 180 ng.

Page 11 of 17



Page 12 of 17

Figure 6: Top: Determined analyte masses versus known, weighed mass for guaiacol (A) and DMDS (B). Bottom: Percent recoveries at different calibration levels. The vertical lines indicate the calculated LOQ.

The measurement range of the method starts at dozens of nanograms and reaches at least several micrograms. Since absolute masses are measured and the size of the headspace vials allows variable sample amounts, the sample size can be adjusted to increase sensitivity or to reduce the signal to stay in the linear range of the detector. In theory, sampling temperature and sampling time could be optimized to target parameters, for example minimal analysis time or full depletion of a certain target analyte in the first extraction. However, one must be aware that lower temperatures and shorter extractions will impair sensitivity and require more extraction steps for reliable quantitation. Higher temperatures might cause degradation of the sample material and reduce the loading capacity of the fiber material. Using a thinner fiber coating (75 instead of 85 µm) was found to have no effect on the extraction, which indicates that the chosen dimension are sufficient to avoid fiber overloading. It is absolutely necessary to keep the set parameters constant, otherwise there will be no correlation between measurement and calibration series. And for a meaningful quantitative evaluation, the detector signal must not be overloaded, and peak areas after depletion of the target analyte must not be included in the calculation of the total peak area.

Quantification of Guaiacol and DMDS in Lignin Samples

The method was applied to four industrial lignins (see Table 1). The content of guaiacol in sample II stands out with its very high amount, which is two orders of magnitude higher than in the other samples. Indeed, Page 12 of 17


Page 13 of 17 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


the subjectively experienced odor of this lignin was much more intense compared to the other lignins, which confirms this result. This sample was the only one stored in a refrigerator and kept for only a few month, which effectively prevented the evaporation of volatile compounds. In contrast, the guaiacol concentrations of sample I, III and IV, all stored at ambient conditions, are in the low ppm region. Comparing these values with published ones by Kalliola et al.,4 they appear to be rather low. However, it has to be taken into account that guaiacol emissions in the cited study were measured at elevated temperatures of 150°C and 190°C. It was also shown,4,7 that increasing the temperature increases the amount of volatiles released from lignin. Hence, we consider these results reasonable. For DMDS, only in sample II were found significant amounts of this volatile. Compared to a value published in literature (0.25 ppm at 60°C),7 it is quite high. Yet again, our finding matches this lignins’s pungent odor and fresh state and can therefore be considered plausible. The absence of DMDS in samples III and IV is not surprising, since the corresponding cooking processes do not employ reduced sulfur reagents, which are the main source of DMDS.1,29 However for sample I, Indulin, it is certainly remarkable that DMDS was completely lost over the years, although it is a kraft lignin and this process relies on Na2S as one of the major pulping chemicals leading to the formation of a range of volatile sulfur compounds including DMDS.29 Kalliola et al.4 reported significant DMDS levels in commercial kraft lignins only at 190°C, while 40°C were sufficient for detection with the present method. All in all, these first results for volatiles in real lignins are consistent with published findings and even augment them. This demonstrates the usefulness of the MHS‐SPME method. By applying it to further analytes, the understanding of volatile release by lignins will be substantially deepened. Table 1: Volatile concentrations of the different lignin samples (n=3). For Samples I, III and IV no distinct DMDS peak could be observed. Sample Sample I (Indulin) Sample II (technical kraft) Sample III (lignosulfonate) Sample IV (organosolv)

Concentration Guaiacol [ppm] 2.3 1234.4 0.5 0.4

Concentration DMDS [ppm] ‐ 33.9 ‐ ‐

Conclusion Gas chromatography with mass spectroscopic detection is very well suited for the qualitative analysis of lignin volatiles. Since samples are taken from the gas volume over the lignin sample, interfering matrix components, for example, salts, water and polymers, are inherently separated and do not interfere with the chromatographic analysis. Initial attempts with a headspace‐sampler to withdraw a gas sample directly from the vial resulted in unacceptably low sensitivity and were therefore dismissed. Using SPME as an extractive sampling method, the necessary sensitivity could be achieved. The method could readily be extended to quantify absolute masses of the most prominent volatiles – guaiacol and dimethyl disulfide – with LODs in the nanogram range. The calibrated measurement range extended into the low microgram range and can be extended by adjusting the weighed sample amount. For reliable results, one must observe that the detected signals are not in the saturated regime of the detector or the fiber and all analyses after the complete extraction of the target analyte are excluded from the quantification. Sample preparation is limited to weighing the sample into a headspace vial, not requiring any solvents, consumables or auxiliaries. The analysis including the data evaluation was fully automatized. The method can be applied to lignin arising from different processes; kraft, sulfite and organosolv were tested. Page 13 of 17



Dimethyl disulfide could only be detected in one of these samples, and the concentration of guaiacol ranged from 0.4 ppm to 1200 ppm. With this tool at hand, volatiles arising from lignin and other biomass can be monitored reliably, paving the way for a deeper understanding of the released substances and their release mechanisms.

Supporting information Comparison of peak shapes of acetic acid and guaiacol, effects of too high sample amount on signal saturation and MHS evaluation, qualitative overview over volatiles found in representative lignins

Acknowledgements We are grateful for the support by our industry partners in the frame of the Flippr2 project, Mondi, Sappi, Zellstoff Pöls AG, a member of heinzel pulp, and Papierholz Austria. The K‐Project Flippr² is funded as part of COMET ‐ Competence Centers for Excellent Technologies promoted by BMVIT, BMWFJ, Styria and Carinthia. The COMET program is managed by FFG.

References (1)

Saake, B.; Lehnen, R. Lignin. In Ullmann’s Encyclopedia of Industrial Chemistry; Wiley‐VCH Verlag GmbH & Co. KGaA, Ed.; Wiley‐VCH Verlag GmbH & Co. KGaA: Weinheim, Germany, 2007. https://doi.org/10.1002/14356007.a15_305.pub3. (2) Stewart, D. Lignin as a Base Material for Materials Applications: Chemistry, Application and Economics. Ind. Crops Prod. 2008, 27 (2), 202–207. https://doi.org/10.1016/j.indcrop.2007.07.008. (3) Naegele, H.; Pfitzer, J.; Ziegler, L.; Inone‐Kauffmann, E. R.; Eisenreich, N. Applications of Lignin Materials and Their Composites (Lignin Applications in Various Industrial Sectors, Future Trends of Lignin and Their Composites). In Lignin in Polymer Composites; Elsevier, 2016; pp 233–244. https://doi.org/10.1016/B978‐0‐323‐35565‐0.00013‐8. (4) Kalliola, A. K.; Savolainen, A.; Ohra‐aho, T.; Faccio, G.; Tamminen, T. Reducing the Content of VOCs of Softwood Kraft Lignins for Material Applications. BioResources 2012, 7 (3), 2871–2882. https://doi.org/10.15376/biores.7.3.2871‐2882. (5) Rocha, S. M.; Gonçalves, V.; Evtuguin, D.; Delgadillo, I. Distinction and Identification of Lignins Based on Their Volatile Headspace Composition. Talanta 2008, 75 (2), 594–597. https://doi.org/10.1016/j.talanta.2007.11.018. (6) Voeller, K.; Bílek, H.; Kreft, J.; Dostálková, A.; Kozliak, E.; Kubátová, A. Thermal Carbon Analysis Enabling Comprehensive Characterization of Lignin and Its Degradation Products. ACS Sustain. Chem. Eng. 2017, 5 (11), 10334–10341. https://doi.org/10.1021/acssuschemeng.7b02392. (7) Kouisni, L.; Gagné, A.; Maki, K.; Holt‐Hindle, P.; Paleologou, M. LignoForce System for the Recovery of Lignin from Black Liquor: Feedstock Options, Odor Profile, and Product Characterization. ACS Sustain. Chem. Eng. 2016, 4 (10), 5152–5159. https://doi.org/10.1021/acssuschemeng.6b00907. (8) Sallem‐Idrissi, N.; Vanderghem, C.; Pacary, T.; Richel, A.; Debecker, D. P.; Devaux, J.; Sclavons, M. Lignin Degradation and Stability: Volatile Organic Compounds (VOCs) Analysis throughout Processing. Polym. Degrad. Stab. 2016, 130, 30–37. https://doi.org/10.1016/j.polymdegradstab.2016.05.028. (9) Hakkarainen, M. Developments in Multiple Headspace Extraction. J. Biochem. Biophys. Methods 2007, 70 (2), 229–233. https://doi.org/10.1016/j.jbbm.2006.08.012. (10) Kolb, B.; Ettre, L. S. Theory and Practice of Multiple Headspace Extraction. Chromatographia 1991, 32 (11), 505–513.

Page 14 of 17


Page 14 of 17

Page 15 of 17 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


(11) Martínez‐Uruñuela, A.; González‐Sáiz, J. M.; Pizarro, C. Multiple Solid‐Phase Microextraction in a Non‐Equilibrium Situation. J. Chromatogr. A 2005, 1089 (1), 31–38. https://doi.org/10.1016/j.chroma.2005.06.063. (12) Ezquerro, Ó.; Ortiz, G.; Pons, B.; Tena, M. T. Determination of Benzene, Toluene, Ethylbenzene and Xylenes in Soils by Multiple Headspace Solid‐Phase Microextraction. J. Chromatogr. A 2004, 1035 (1), 17–22. https://doi.org/10.1016/j.chroma.2004.02.030. (13) Gröning, M.; Hakkarainen, M. Multiple Headspace Solid‐Phase Microextraction of 2‐Cyclopentyl‐ Cyclopentanone in Polyamide 6.6: Possibilities and Limitations in the Headspace Analysis of Solid Hydrogen‐Bonding Matrices. J. Chromatogr. A 2004, 1052 (1), 61–68. https://doi.org/10.1016/j.chroma.2004.08.112. (14) Ezquerro, Ó.; Tena, M. T. Determination of Odour‐Causing Volatile Organic Compounds in Cork Stoppers by Multiple Headspace Solid‐Phase Microextraction. J. Chromatogr. A 2005, 1068 (2), 201–208. https://doi.org/10.1016/j.chroma.2005.01.089. (15) Carrillo, J. D.; Tena, M. T. Determination of Volatile Compounds in Antioxidant Rosemary Extracts by Multiple Headspace Solid‐Phase Microextraction and Gas Chromatography. Flavour Fragr. J. 2006, 21 (4), 626–633. https://doi.org/10.1002/ffj.1630. (16) Flores, M.; Hernández, D. Optimization of Multiple Headspace Solid‐Phase Microextraction for the Quantification of Volatile Compounds in Dry Fermented Sausages. J. Agric. Food Chem. 2007, 55 (21), 8688–8695. https://doi.org/10.1021/jf0713758. (17) Serrano, E.; Beltrán, J.; Hernández, F. Application of Multiple Headspace‐Solid‐Phase Microextraction Followed by Gas Chromatography–Mass Spectrometry to Quantitative Analysis of Tomato Aroma Components. J. Chromatogr. A 2009, 1216 (1), 127–133. https://doi.org/10.1016/j.chroma.2008.11.026. (18) Costa, R.; Tedone, L.; De Grazia, S.; Dugo, P.; Mondello, L. Multiple Headspace‐Solid‐Phase Microextraction: An Application to Quantification of Mushroom Volatiles. Anal. Chim. Acta 2013, 770, 1–6. https://doi.org/10.1016/j.aca.2013.01.041. (19) San Román, I.; Alonso, M. L.; Bartolomé, L.; Alonso, R. M.; Fañanás, R. Analytical Strategies Based on Multiple Headspace Extraction for the Quantitative Analysis of Aroma Components in Mushrooms. Talanta 2014, 123, 207–217. https://doi.org/10.1016/j.talanta.2014.01.021. (20) Sgorbini, B.; Bicchi, C.; Cagliero, C.; Cordero, C.; Liberto, E.; Rubiolo, P. Herbs and Spices: Characterization and Quantitation of Biologically‐Active Markers for Routine Quality Control by Multiple Headspace Solid‐Phase Microextraction Combined with Separative or Non‐Separative Analysis. J. Chromatogr. A 2015, 1376, 9–17. https://doi.org/10.1016/j.chroma.2014.12.007. (21) Cantú, A. A. Comments on the Multiple Headspace‐Solid Phase Microextraction (MHS‐SPME) Technique for Dating Inks. Forensic Chem. 2017, 3, 14–20. https://doi.org/10.1016/j.forc.2016.10.008. (22) McNaught, A. D.; Wilkinson, A.; Jenkins, A. Retention Index. IUPAC. Compendium of Chemical Terminology, 2nd ed. (the “Gold Book”). XML on‐line corrected version: http://goldbook.iupac.org (2006‐); Nic, M., Jirat, J., Kosata, B., Eds.; Blackwell Scientific Publications: Oxford, 1997. (23) Kim, S.; Thiessen, P. A.; Bolton, E. E.; Chen, J.; Fu, G.; Gindulyte, A.; Han, L.; He, J.; He, S.; Shoemaker, B. A.; et al. PubChem Substance and Compound Databases. Nucleic Acids Res. 2016, 44 (Database issue), D1202–D1213. https://doi.org/10.1093/nar/gkv951. (24) Abraham, J. ICH Harmonised Tripartitite Guideline: Validation of Analytical Procedures: Text and Methodology. In International Conference On Harmonisation Of Technical Requirements For Registration Of Pharmaceuticals For Human Use; Brouder, A., Tietje, C., Eds.; Brill, 2009; pp 1041– 1054. https://doi.org/10.1163/ej.9789004163300.i‐1081.897. (25) Haahr, M. RANDOM.ORG: True Random Number Service https://www.random.org (accessed Jun 1, 2018).

Page 15 of 17



(26) Scanlon, J. T.; Willis, D. E. Calculation of Flame Ionization Detector Relative Response Factors Using the Effective Carbon Number Concept. J. Chromatogr. Sci. 1985, 23 (8), 333–340. https://doi.org/10.1093/chromsci/23.8.333. (27) de Saint Laumer, J.‐Y.; Leocata, S.; Tissot, E.; Baroux, L.; Kampf, D. M.; Merle, P.; Boschung, A.; Seyfried, M.; Chaintreau, A. Prediction of Response Factors for Gas Chromatography with Flame Ionization Detection: Algorithm Improvement, Extension to Silylated Compounds, and Application to the Quantification of Metabolites: Gas Chromatography. J. Sep. Sci. 2015, 38 (18), 3209–3217. https://doi.org/10.1002/jssc.201500106. (28) Tena, M. T.; Carrillo, J. D. Multiple Solid‐Phase Microextraction: Theory and Applications. TrAC Trends Anal. Chem. 2007, 26 (3), 206–214. https://doi.org/10.1016/j.trac.2007.01.008. (29) Karnofski, M. A. Odor Generation in the Kraft Process. J Chem Educ 1975, 52 (8), 490–492.

Page 16 of 17


Page 16 of 17

Page 17 of 17 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Synopsis Odorous substances emitted from lignin, a promising raw material for renewable products, were identified and quantified by a GC‐MS method requiring no sample preparation.

For Table of Contents Use Only

Page 17 of 17


Quantification of Volatiles from Technical Lignins by Multiple

Recommend Documents